From ae260060fddaaeccf2b68a1de51d5e780b099dfb Mon Sep 17 00:00:00 2001
From: Volpeon <git@volpeon.ink>
Date: Tue, 27 Sep 2022 10:54:40 +0200
Subject: Undo textual inversion dataset improvements

---
 data/textual_inversion/csv.py | 31 +++++++++++++++++++++----------
 1 file changed, 21 insertions(+), 10 deletions(-)

(limited to 'data')

diff --git a/data/textual_inversion/csv.py b/data/textual_inversion/csv.py
index 38ffb6f..0d1e96e 100644
--- a/data/textual_inversion/csv.py
+++ b/data/textual_inversion/csv.py
@@ -80,19 +80,14 @@ class CSVDataset(Dataset):
 
         self.placeholder_token = placeholder_token
 
+        self.size = size
+        self.center_crop = center_crop
         self.interpolation = {"linear": PIL.Image.LINEAR,
                               "bilinear": PIL.Image.BILINEAR,
                               "bicubic": PIL.Image.BICUBIC,
                               "lanczos": PIL.Image.LANCZOS,
                               }[interpolation]
-        self.image_transforms = transforms.Compose(
-            [
-                transforms.Resize(size, interpolation=transforms.InterpolationMode.BILINEAR),
-                transforms.CenterCrop(size) if center_crop else transforms.RandomCrop(size),
-                transforms.ToTensor(),
-                transforms.Normalize([0.5], [0.5]),
-            ]
-        )
+        self.flip = transforms.RandomHorizontalFlip(p=flip_p)
 
         self.cache = {}
 
@@ -107,9 +102,9 @@ class CSVDataset(Dataset):
 
         example = {}
         image = Image.open(image_path)
+
         if not image.mode == "RGB":
             image = image.convert("RGB")
-        image = self.image_transforms(image)
 
         text = text.format(self.placeholder_token)
 
@@ -122,8 +117,24 @@ class CSVDataset(Dataset):
             return_tensors="pt",
         ).input_ids[0]
 
+        # default to score-sde preprocessing
+        img = np.array(image).astype(np.uint8)
+
+        if self.center_crop:
+            crop = min(img.shape[0], img.shape[1])
+            h, w, = img.shape[0], img.shape[1]
+            img = img[(h - crop) // 2:(h + crop) // 2,
+                      (w - crop) // 2:(w + crop) // 2]
+
+        image = Image.fromarray(img)
+        image = image.resize((self.size, self.size),
+                             resample=self.interpolation)
+        image = self.flip(image)
+        image = np.array(image).astype(np.uint8)
+        image = (image / 127.5 - 1.0).astype(np.float32)
+
         example["key"] = "-".join([image_path, "-", str(flipped)])
-        example["pixel_values"] = image
+        example["pixel_values"] = torch.from_numpy(image).permute(2, 0, 1)
 
         self.cache[image_path] = example
         return example
-- 
cgit v1.2.3-54-g00ecf